MEWS: Artificial intelligence for the detection of fake news on the Internet

Disinformation, perpetrated through false or distorted news, is one of the great vices of the Internet today, in which social networks have acted as a catalyst for this problem.

A recently published research article proposes a system for analyzing and detecting manipulation of social networks in real time.

An AI system to filter fake news

Researchers from the Universities of Notre Dame and Louisville, presented the beta version of MEWS (Misinformation Early Warning System). It is a system that writes the various aspects of ingestion, tampering detection, and graphics algorithms used to determine, in near real time, the relationships between social media images used to convey news information, as that these expressions arise and spread on social media platforms.

By combining these various technologies into a single processing pipeline, MEWS gains the ability to identify manipulated media elements.

Under this logic of analysis, the appearance of a new content manipulation, followed by a rapid spread of the manipulated material, suggests a disinformation campaign.

The proposed system uses a database of images and videos from various social media platforms in its tracking. This sample is processed by several state-of-the-art AI systems, capable of identifying manipulations and certain key elements, such as the faces that appear, visible objects, texts and other visual characteristics.

Then, after detecting that, the system builds a media graph that pairs sub-images, objects and similar manipulations, to display them in a comparative way in an interactive, easily navigable and searchable user interface.

Various demonstrations, documented in this study report, showed that the proposed system can interactively reveal emerging trends in social media images in near real time and identify repeated manipulations and alterations in media elements and platforms.

For civilian users, MEWS will be available in the future to be used as a search interface, to understand how disinformation spreads on social networks, in the midst of these times when its social influence is really tremendous.